Hydrothermal flow paths in the upper oceanic crust: magnitude and distribution of porosity and permeability in altered, fractured lavas, Oman ophiolite

Brett, Alannah C. (2021). Hydrothermal flow paths in the upper oceanic crust: magnitude and distribution of porosity and permeability in altered, fractured lavas, Oman ophiolite (Unpublished). (Dissertation, Institute of Geological Sciences, Science)

[img]
Preview
Text (PhD Thesis)
PhD_Thesis_Alannah_C._Brett_2021.pdf - Other
Available under License BORIS Standard License.

Download (22MB) | Preview

Convection of seawater and consequent fluid–rock interaction in the basaltic upper crust is responsible for numerous important processes, including the formation of valuable ore deposits at the seafloor, the buffering of composition of the oceans, and the cooling of the oceanic lithosphere. Circulation of hydrothermal fluid is facilitated by the porosity and permeability of the sheeted dyke complexes and submarine lavas that make up the upper crust. Previous studies have assessed the porosity and permeability of fresh basalts and basalts altered at low temperatures to clay and zeolite assemblages, but little is known about the hydraulic properties of the rocks deeper in the crust, where greenschist-facies dominate. Knowledge of these properties is a key prerequisite to performing realistic numerical simulations of the coupled processes caused by hydrothermal circulation. This Thesis provides new perspectives on the porosity and permeability architecture of the upper oceanic crust where it is altered pervasively to greenschist-facies assemblages, by investigating the MORB-type axial lavas and underlying sheeted dyke complex of the Semail ophiolite, Oman. The exceptional outcrop exposures of largely undeformed oceanic lithosphere in Oman provide a well-characterized context for observations and mapping of petrophysical properties at centimetre- to 100-meter scales. Sampling and mapping distinguished rocks altered by downwelling hydrothermal fluids (chl+alb+act-bearing spilites) from those formed by upwelling fluids (epi+qz-bearing epidosites). These observations can be upscaled numerically to construct a model of rock-matrix and fracture-network flow paths that represents the upper crust at the km-scale. Development and interpretation of this model by integrating new measurements of rock-matrix porosity and permeability, fracture maps, high P–T experiments and flow simulations are the principle goals of this Thesis.

The first part of this Thesis assesses whether standard permeability measurements of spilites and epidosites at room conditions represent in-situ permeabilities at the effective P (5–55 MPa) and T (25–450 °C) of hydrothermal alteration. Our experiments in a Paterson permeameter at 50 MPa show that heating from 25 to 450 °C reduces permeability by ~40–50% and increasing pressure at 25 °C from 5 to 55 MPa reduces permeability by ~70–80%. Based on these results we provide quantitative correlations to adjust results of standard measurements to their in-situ, high P–T values. However, duplicate measurements demonstrate that hand-sample heterogeneity spans up to three orders of magnitude in permeability, outweighing the P–T effects and rendering corrections unnecessary for large natural datasets. Consequently, we performed all further permeability measurements reported in this Thesis at high P without adjustment for elevated temperature.

The second study in the Thesis examines how rock-matrix porosity and permeability of spilites and epidosites depends on volcanic rock type and on the extent of the spilite-to-epidosite alteration reaction. Our results demonstrate that the porosity and permeability of spilites increase in the order dykes < massive lava flows < pillow lava rims < pillow lava cores < interpillow hyaloclastites. These findings are consistent with field observations showing how epidosite alteration progresses through the different lava types. Additionally, combination of our data on pillow cores, rims and interpillow hyaloclastites reveals that spilite pillow stacks have permeabilities up to ~2 × 10−16 m2. This relatively high value is an important step towards reconciling the large differences between the low rock-matrix permeabilities determined in laboratory measurements and those required by numerical models constrained to reproduce observations of heat and fluid discharge from the crust.

As part of this study we take a more in-depth examination of epidosites. Although epidosites are thought to form mainly at the base of the sheeted dyke complex, numerous occurrences up to ~1 km2 in outcrop have been documented in the Semail lavas. It is known that extreme fluid–rock ratios are required to convert the precursor spilites to endmember epidosites, but mechanism by which the hot (230–450 °C) fluids become so focused during upflow, apparently independently of any fracture networks, is still unclear. Previous studies have suggested epidotising fluids may create their own porosity and permeability, enabling self-propagation without fracture permeability. We quantified this process by directly comparing the porosity, permeability and mineralogy of pairs of spatially adjacent spilite and epidosite samples, and we integrated these results in reactive-transport simulations, showing that the spilite-to-epidosite reaction creates 9–14 vol.% porosity and up to 4.5 order of magnitude (~10−15 m2) permeability. This results in a highly porous (13–26 vol.%) and permeable (~10−15 m2) epidosite rock-matrix, which focuses hydrothermal upflow in the absence of significant fracture networks.

In the third study, we bring together upscaled rock-matrix properties for lava stacks with fracture distribution maps, to create 100-meter-scale, dual-porosity, dual-permeability models in order to estimate the bulk permeability of the volcanic layer of the upper crust. The current view is that regional fluid flow in the lavas is characterised by fractured aquifer properties, dominated by fracture networks and extensional faults. However, between kilometre-scale faults and their localised damage zones, termed ‘distal zones’, the distribution and connectivity of fracture networks are not quantified. Moreover, the results from our rock-matrix study and the pervasive alteration of the rock-matrix in ophiolites demonstrates that rock-matrix pathways cannot be ignored at a regional scale. Combining rock-matrix porosities and permeabilities at outcrop scale (weighted for pillow core, pillow rim and interpillow hyaloclastite averages from our previous study) yields porosities of ~10–12 vol.% and permeabilities of ~2.5 × 10−16 m2 for spilite pillow stacks. In distal zones, the fractures (mapped as hydrothermal veins) are sparse and discrete, with fracture intensities of only ~0.005 m of fracture per m2 of outcrop, yet an order of magnitude higher intensity is present in damage zones (~0.063). Using dfnWorks software, simulations of coupled fracture- and matrix-flow through upscaled discrete fracture matrix models integrate these properties and confirm that damage zones are indeed fracture-controlled fluid channels with permeabilities of ~10−11 m2. This contrasts with the permeabilities of ~5 × 10−16 m2 found for sparsely fractured distal zones, in which flow is controlled by the rock-matrix and in which fractures contribute only ~0.25 of an order of magnitude to the bulk permeability. This supports the idea that spilitising and epidotising fluids largely migrate through the rock-matrix in distal zones of the upper oceanic crust.

The results of this Thesis constitute a new perspective on hydrothermal flow paths in the upper oceanic crust. The new datasets of spilite and epidosite rock-matrix petrophysical properties, accounting for in-situ pressure and temperature, fill gaps in previous understanding and complement available data on fresh and low-temperature altered basalts. The changes in porosity and permeability during the spilite-to-epidosite reaction reveals how hydrothermal upflow becomes self-focused and attain extreme water–rock ratios. While we confirmed that fault zones spaced at 0.5–1 km distances act as major conduits, a new insight is that the large blocks of crust between such faults behave as permeable rock-matrix aquifers, which facilitate the pore-scale fluid–rock interactions necessary to extract metals and heat for formation of VMS deposits and other chemical elements and isotopes that buffer the composition of the oceans. The next generation of numerical thermal–hydraulic–chemical models will be able to incorporate the km-scale permeability architecture provided by our analysis of hydrothermally altered distal zones and fault-damage zones, to better describe the water–rock interactions in the upper crust.

Item Type:

Thesis (Dissertation)

Division/Institute:

08 Faculty of Science > Institute of Geological Sciences

UniBE Contributor:

Brett, Alannah Charlotte, Diamond, Larryn William, Herwegh, Marco

Subjects:

500 Science > 550 Earth sciences & geology

Funders:

Organisations 169653 not found.; Organisations 188567 not found.

Language:

English

Submitter:

Alannah Charlotte Brett

Date Deposited:

23 Jan 2023 11:02

Last Modified:

13 Aug 2023 02:04

BORIS DOI:

10.48350/177414

URI:

https://boris.unibe.ch/id/eprint/177414

Actions (login required)

Edit item Edit item
Provide Feedback